Efficacy of hematopoietic stem cell mobilization regimens in patients with hematological malignancies: a systematic review and network meta-analysis of randomized controlled trials

Abstract

Background: Efficient mobilization of hematopoietic stem cells (HSCs) from bone marrow niche into circulation is the key to successful collection and transplantation in patients with hematological malignancies. The efficacy of various HSCs mobilization regimens has been widely investigated, but the results are inconsistent.

Methods: We performed comprehensive databases searching for eligible randomized controlled trials (RCTs) that comparing the efficacy of HSCs mobilization regimens in patients with hematological malignancies. Bayesian network meta-analyses were performed with WinBUGS. Standard dose of granulocyte colony-stimulating factor (G-CSF SD) was chosen as the common comparator. Estimates of relative treatment effects for other regimens were reported as mean differences (MD) or odds ratio (OR) with associated 95% credibility interval (95% CrI). The surface under the cumulative ranking curve (SUCRA) were obtained to present rank probabilities of all included regimens.

Results: Databases searching and study selection identified 44 eligible RCTs, of which the mobilization results are summarized. Then we compared the efficacy of mobilization regimens separately for patients with multiple myeloma (MM) and non-Hodgkin lymphoma (NHL) by including 13 eligible trials for network meta-analysis, involving 638 patients with MM and 592 patients with NHL. For patients with MM, data are pooled from 8 trials for 6 regimens, including G-CSF in standard dose (SD) or reduced dose (RD) combined with cyclophosphamide (CY), intermediate-dose cytarabine (ID-AraC) or plerixafor. The results show that compared with G-CSF SD alone, 3 regimens including ID-AraC + G-CSF SD (MD 14.29, 95% CrI 9.99-18.53; SUCRA 1.00), G-CSF SD + Plerixafor SD (MD 4.15, 95% CrI 2.92-5.39; SUCRA 0.80), and CY + G-CSF RD (MD 1.18, 95% CrI 0.29-2.07; SUCRA 0.60) are associated with significantly increased total number of collected CD34⁺ cells (× 10⁶/kg), among which ID-AraC + G-CSF SD ranked first with a probability of being best regimen of 100%. Moreover, ID-AraC + G-CSF SD and G-CSF SD + Plerixafor SD are associated with significantly higher successful rate of achieving optimal target (collecting ≥ 4-6 × 10⁶ CD34⁺ cells/kg). For patients with NHL, data are pooled from 5 trials for 4 regimens, the results show that compared with G-CSF SD alone, G-CSF SD + Plerixafor SD (MD 3.62, 95% CrI 2.86-4.38; SUCRA 0.81) and G-CSF SD plus the new CXC chemokine receptor-4 (CXCR-4) antagonist YF-H-2015005 (MD 3.43, 95% CrI 2.51-4.35; SUCRA 0.69) are associated with significantly higher number of total CD34⁺ cells collected. These 2 regimens are also associated with significantly higher successful rate of achieving optimal target. There are no significant differences in rate of achieving optimal target between G-CSF SD + Plerixafor SD and G-CSF + YF-H-2015005.

Conclusions: In conclusion, ID-AraC plus G-CSF is associated with the highest probability of being best mobilization regimen in patients with MM. For patients with NHL, G-CSF in combination with plerixafor or YF-H-2015005 showed similar improvements in HSCs mobilization efficacy. The relative effects of other chemotherapy-based mobilization regimens still require to be determined with further investigations.

Keywords: Cyclophosphamide; G-CSF; Hematological malignancies; Hematopoietic stem cell mobilization; Plerixafor.

Fig. 1

PRISMA flowchart of study selection

Fig. 2

Network plot for total number of collected CD34⁺ cells. Network plot depicting all direct comparisons in included trials with data about the total number of collected CD34⁺ cell (× 10⁶/kg) for patients with MM (A) and NHL (B). Each node represents a mobilization regimen, while each line represents direct comparison between two regimens, with the thickness reflecting the number of times of direct comparisons. CY, cyclophosphamide; G-CSF, granulocyte colony-stimulating factor; ID-AraC, intermediate-dose cytarabine; RD, reduced dose; SD, standard dose; YF-H-2015005, a new CXCR4 antagonist

Fig. 3

Forest plot of meta-analysis results for the total number of collected CD34⁺ cells. Forest plot regarding the network meta-analysis results of the total number of collected CD34⁺ cells (× 10⁶/kg) for patients with MM (A) and NHL (B). G-CSF SD is the common comparator. Estimate of relative treatment effect for other mobilization regimens are reported as mean differences (MD) with the associated 95% credibility interval (95% CrI). CY, cyclophosphamide; G-CSF, granulocyte colony-stimulating factor; ID-AraC, intermediate-dose cytarabine; RD, reduced dose; SUCRA, surface under the cumulative ranking curve; SD, standard dose; YF-H-2015005, a new CXCR4 antagonist

Fig. 4

Pooled results for successful rate of achieving optimal target. Pooled ORs and 95% CrI for all possible head-to-head comparisons regarding the successful rate of achieving optimal target for patients with MM (A) and NHL (B). OR, odds ratio. 95% CrI, 95% credibility interval. CY, cyclophosphamide; FD, fixed dose; G-CSF, granulocyte colony-stimulating factor; ID-AraC, intermediate-dose cytarabine; RD, reduced dose; SD, standard dose; YF-H-2015005, a new CXCR4 antagonist